1: % ****** Start of file apssamp.tex ******
2: % This file is part of the APS files in the REVTeX 4 distribution.
3: % Version 4.0 of REVTeX, August 2001
4: % Copyright (c) 2001 The American Physical Society.
5: % See the REVTeX 4 README file for restrictions and more information.
6: % TeX'ing this file requires that you have AMS-LaTeX 2.0 installed
7: % as well as the rest of the prerequisites for REVTeX 4.0
8: % See the REVTeX 4 README file
9: % It also requires running BibTeX. The commands are as follows:
10: % 1) latex apssamp.tex
11: % 2) bibtex apssamp
12: % 3) latex apssamp.tex
13: % 4) latex apssamp.tex
14: %\documentclass[preprint,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
15: % Some other (several out of many) possibilities
16: %\documentclass[preprint,aps]{revtex4}
17: %\documentclass[preprint,aps,draft]{revtex4}
18: %\documentclass[prb]{revtex4}% Physical Review B
19: % Include figure files
20: % Align table columns on decimal point
21: % bold math
22: %\nofiles
23:
24:
25: \documentclass[showpacs,preprintnumbers,onecolumn,12pt]{revtex4}
26: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
27: \usepackage{amssymb}
28: \usepackage{amsfonts}
29: \usepackage{amsmath}
30: \usepackage{graphicx}
31: \usepackage{dcolumn}
32: \usepackage{bm}
33:
34: \setcounter{MaxMatrixCols}{10}
35: %TCIDATA{OutputFilter=LATEX.DLL}
36: %TCIDATA{Version=5.00.0.2557}
37: %TCIDATA{<META NAME="SaveForMode" CONTENT="1">}
38: %TCIDATA{LastRevised=Tuesday, May 30, 2006 12:20:57}
39: %TCIDATA{<META NAME="GraphicsSave" CONTENT="32">}
40: %TCIDATA{Language=American English}
41:
42: \input{tcilatex}
43:
44: \begin{document}
45:
46: \title{Homoclinic Signatures of Dynamical Localization}
47: \author{Ricardo Chac\'{o}n}
48: \affiliation{Departamento de Electr\'{o}nica e Ingenier\'{\i}a Electromec\'{a}nica,
49: Escuela de Ingenier\'{\i}as Industriales, Universidad de Extremadura,
50: Apartado Postal 382, E-06071 Badajoz, Spain}
51: \date{\today}
52:
53: \begin{abstract}
54: It is demonstrated that the oscillations in the width of the momentum
55: distribution of atoms moving in a phase-modulated standing light field, as a
56: function of the modulation amplitude $\lambda $, are correlated with the
57: variation of the chaotic layer width in energy of an underlying \textit{%
58: effective} pendulum. The maximum effect of dynamical localization and the
59: nearly perfect delocalization are associated with the maxima and minima,
60: respectively, of the chaotic layer width. It is also demonstrated that
61: kinetic energy is conserved as an almost adiabatic invariant at the minima
62: of the chaotic layer width, and that the system is \textit{accurately}
63: described by delta-kicked rotors at sufficiently large zeros of the Bessel
64: functions $J_{0}\left( \lambda \right) $ and $J_{1}\left( \lambda \right) $.
65: Numerical calculations of kinetic energy and Lyapunov exponents confirm all
66: the theoretical predictions.
67: \end{abstract}
68:
69: \pacs{05.45.Mt, 42.50.Vk}
70: \maketitle
71:
72: % Force line breaks with \\
73:
74: %Lines break automatically or can be forced with \\
75:
76: % It is always \today, today,
77: % but any date may be explicitly specified
78:
79: % PACS, the Physics and Astronomy
80: % Classification Scheme.
81: %\keywords{Suggested keywords}%Use showkeys class option if keyword
82: %display desired
83:
84: Nonautonomous Hamiltonian systems exhibit a wide range of novel effects with
85: different manifestations in the quantum and classical domains [1,2]. A
86: situation of especial interest occurs when the classical dynamics undergoes
87: a transition from stability to chaos as a control parameter is varied [3].
88: In this regard, dynamical localization (DL) is a specifically quantum
89: phenomenon taking place in systems where the classical dynamics exhibits
90: chaotic diffusion in momentum space, while the asymptotic quantum momentum
91: distribution freezes to a steady state although it is initially similar to
92: its classical counterpart. This phenomenon is attributed to quantum
93: interferences among the diffusive paths which on average over long times are
94: completely destructive. These interferences giving rise to chaotic diffusion
95: are properly described in terms of gradual dephasing of the various Floquet
96: states that contribute to the initial state. After a characteristic time
97: (the so-called Ehrenfest time $\tau _{E}\sim \ln \left( S/\hbar \right)
98: /\Lambda ^{+}$, where $\Lambda ^{+}$ is the sum of positive Lyapunov
99: exponents (LEs) and $S$ is a typical classical action [4]), the momentum
100: distribution covers all significantly populated Floquet states as these
101: become completely dephased, leading to the freezing of the diffusive growth.
102: This phenomenon has been discussed in model systems such as the kicked
103: rotator [5] as well as in the center-of-mass motion of a single ion confined
104: in a Paul trap and interacting with a laser field [6]. While no classical or
105: semi-classical explanation for DL has as yet been proposed, a formal analogy
106: with the Anderson localization appearing in disordered systems has been
107: discussed [7]: DL occurs over the time coordinate with localization in
108: momentum space, while Anderson localization refers to spatial coordinates.
109: Additional evidence of the DL phenomenon has been provided by the momentum
110: transfer from a modulated standing light wave to a sample of ultracold atoms
111: [8-12]. It was found experimentally [9,11] and numerically [10,11] that the
112: width of the atomic momentum distribution exhibits oscillations as a
113: function of the modulation amplitude, while a theoretical analysis [8]
114: predicted its overall monotonous decrease for sufficiently large modulation
115: amplitudes (\textit{quantum} regime). Also, a preliminary qualitative
116: analysis of such oscillatory behavior was previously discussed in terms of
117: the zeros of Bessel functions [10,11]. However, no theoretical approach has
118: as yet been proposed to explain in detail the waveform of the aforementioned
119: oscillations or the amplitude values at which the maximum effect of DL had
120: been found experimentally and numerically. In this Letter, such fundamental
121: questions in terms of the strength of \textit{homoclinic instabilities} are
122: explained by studying the dependence on the modulation amplitude of the
123: width in energy of the chaotic separatrix layer associated with an
124: underlying effective perturbed pendulum. Remarkably, homoclinic
125: instabilities \textit{also} imprint a clear signature in the quantum spectra
126: of a particle confined in a stadium billiard [13].
127:
128: \textit{Effective model.}$-$The motion of an atom in a phase-modulated light
129: field, as created for instance by an oscillating mirror, is described [8] by
130: the Hamiltonian $\widetilde{H}=\widetilde{p}^{2}/(2M)-V_{0}\cos \left[ 2k%
131: \widetilde{x}-\lambda \sin \left( \omega t\right) \right] $, where $M,%
132: \widetilde{x},\widetilde{p},V_{0},k,\lambda ,$ and $\omega $ denote the
133: mass, position, momentum, height of the periodic potential, dimensionless
134: modulation depth, wave number, and frequency modulation, respectively.
135: Switching to scaled dimensionless variables $\tau \equiv \omega t,x\equiv 2k%
136: \widetilde{x},p\equiv \left[ 2k/\left( M\omega \right) \right] \widetilde{p}$%
137: , $H\equiv \left[ 4k^{2}/\left( M\omega ^{2}\right) \right] \widetilde{H}$,
138: one obtains the dimensionless Hamiltonian%
139: \begin{equation}
140: H=\frac{p^{2}}{2}-\kappa \cos \left( x-\lambda \sin \tau \right) , \tag{1}
141: \end{equation}%
142: where $x,p,\tau ,$ and $\kappa $ are the dimensionless coordinate, momentum,
143: time, and height of the potential, respectively. It is worth noting that the
144: Hamiltonian (1) also describes the dynamics of a small dissipationless
145: Josephson junction driven by an external sinusoidal current source [14]. To
146: discuss an energy-based characterization of the chaotic dynamics for $%
147: \lambda >0$, one rewrites Eq. (1) as $H=H_{0}+H_{1}$ where $H_{0}\equiv
148: p^{2}/2-\kappa \cos x$, $H_{1}\equiv \kappa \left[ \cos x-\cos \left(
149: x-\lambda \sin \tau \right) \right] $. Now one assumes that $H$ may be
150: regarded as the Hamiltonian of an effective perturbed pendulum, with $H_{1}$
151: being the Hamiltonian perturbation for $\lambda >0$. The change of energy of
152: the unperturbed pendulum because of the influence of $H_{1}$ can be
153: determined from $dH_{0}/d\tau =\kappa p[\sin x-\sin \left( x-\lambda \sin
154: \tau \right) ]$ and hence the change of kinetic energy: $d\left(
155: p^{2}/2\right) /d\tau =-\kappa p\sin \left( x-\lambda \sin \tau \right) $.
156: Since homoclinic dynamics is the origin of the chaotic behavior in
157: Hamiltonian systems, it is useful to calculate the change of the averaged
158: energy along the separatrix $x_{0,\pm }\left( \tau \right) =\pm 2\arctan
159: \left\{ \sinh \left[ \sqrt{\kappa }\left( \tau -\tau _{0}\right) \right]
160: \right\} $, $p_{0,\pm }\left( \tau \right) =\pm 2\sqrt{\kappa }\func{sech}%
161: \left[ \sqrt{\kappa }\left( \tau -\tau _{0}\right) \right] $ of the
162: unperturbed pendulum $H_{0}$: $\left\langle dH_{0}/d\tau \right\rangle
163: \equiv \int_{-\infty }^{\infty }\left( dH_{0}/d\tau \right) d\tau $, where $%
164: \tau =\tau _{0}$ gives the centre of the soliton [2]. Expanding the term $%
165: \sin \left( x-\lambda \sin \tau \right) $ by using the relationship $\exp
166: \left( i\lambda \sin \tau \right) =\sum_{n}J_{n}\left( \lambda \right) \exp
167: \left( in\tau \right) $ of the Bessel functions, one straightforwardly finds%
168: \begin{equation}
169: \left\langle \frac{dH_{0}}{d\tau }\right\rangle =\left\langle \frac{d\left(
170: p^{2}/2\right) }{d\tau }\right\rangle \approx d\left( \lambda ,\kappa
171: \right) \equiv \sum_{n=1}^{\infty }\left[ a_{n}^{2}\left( \lambda ,\kappa
172: \right) +2b_{n}^{2}\left( \lambda ,\kappa \right) +2a_{n}\left( \lambda
173: ,\kappa \right) b_{n}\left( \lambda ,\kappa \right) \right] ^{1/2}, \tag{2}
174: \end{equation}%
175: where $a_{n}\left( \lambda ,\kappa \right) \equiv 16\pi n^{2}J_{2n}\left(
176: \lambda \right) J_{0}^{-2}\left( \lambda \right) \func{csch}\left( n\pi /%
177: \sqrt{\kappa \left\vert J_{0}\left( \lambda \right) \right\vert }\right)
178: /\kappa $, $b_{n}\left( \lambda ,\kappa \right) \equiv 4\pi \left\vert
179: J_{2n-1}\left( \lambda \right) /J_{0}\left( \lambda \right) \right\vert %
180: \left[ 1+2\left( 2n-1\right) ^{2}/\left\vert \kappa J_{0}\left( \lambda
181: \right) \right\vert \right] \func{sech}\left[ \left( 2n-1\right) \pi /(2%
182: \sqrt{\kappa \left\vert J_{0}\left( \lambda \right) \right\vert })\right] $.
183: The width function $d\left( \lambda ,\kappa \right) $ provides an estimate
184: of the width in energy of the chaotic separatrix layer. Figure 1(a) clearly
185: shows that the behavior of the width function $d\left( \lambda \right) $
186: correlates with those of the classical and quantum mechanical momentum
187: distributions obtained in Ref. [10]. In particular, Fig. 1(b) shows that the
188: maxima of $d\left( \lambda \right) $ approximately correspond to the
189: amplitude values where a maximum effect of DL appears, while the minima of $%
190: d\left( \lambda \right) $ correspond to the values where classical and
191: quantum distributions approximately coincide (i.e., DL disappears). Figure 1
192: also shows that the function $d\left( \lambda \right) $ presents two kinds
193: of minima: those exactly occurring at the zeros of $J_{0}\left( \lambda
194: \right) $, where the estimate of the chaotic layer width becomes null, and
195: those approximately occurring at the zeros of $J_{1}\left( \lambda \right) $%
196: , where the chaotic layer exhibits a small but finite width. These two kinds
197: of minima suggest that the kinetic energy could be conserved as an \textit{%
198: almost adiabatic invariant} [15] at the zeros $\left\{ \lambda _{0}^{\ast
199: }\right\} $ of $J_{0}\left( \lambda \right) $ for a \textit{certain }set of
200: initial conditions, and at the zeros $\left\{ \lambda _{1}^{\ast }\right\} $
201: of $J_{1}\left( \lambda \right) $ for another set of initial conditions.
202: Numerical calculations of the maximal LE (see Figs. 2-4) and the kinetic
203: energy (data not shown) confirm these theoretical predictions. Indeed,
204: system (1) presents three LEs $\left( \Lambda ^{+},0,\Lambda ^{-}\right) $
205: which satisfy $\Lambda ^{+}=-$ $\Lambda ^{-}$. When all three LEs vanish,
206: the motion is quasi-periodic and the kinetic energy is conserved in the
207: aforementioned sense, as in the examples shown in Figs. 3 and 4. Remarkably,
208: one finds that the function $d\left( \lambda \right) $ also correlates with $%
209: \Lambda ^{+}$ and hence with $\tau _{E}^{-1}$ (see Fig. 2). Comparison of
210: Figs. 3 and 4 clearly indicates that DL strongly depends on the initial
211: momentum, which should be taken into account in experiments [16].
212:
213: \textit{Dynamics at the zeros of }$J_{0}\left( \lambda \right) $ \textit{and
214: }$J_{1}\left( \lambda \right) $.$-$Using well-known properties of $%
215: J_{n}\left( \lambda \right) $ and $\delta _{T}$ (the Dirac comb of period $T$%
216: ) [17,2], one straightforwardly obtains that Eq. (1) at $\left\{ \lambda
217: _{0}^{\ast }\right\} ,\left\{ \lambda _{1}^{\ast }\right\} $ reduces to%
218: \begin{eqnarray}
219: \frac{d^{2}z}{d\tau ^{\prime \prime 2}}+K_{0}\delta _{T}\left( \tau ^{\prime
220: \prime }\right) \sin z &=&O\left( 1/\lambda _{0}^{\ast 2}\right) ,
221: \TCItag{3} \\
222: \frac{d^{2}x}{d\tau ^{\prime 2}}+K_{1}\delta _{T}\left( \tau ^{\prime
223: }\right) \sin x &=&O[J_{0}\left( \lambda _{1}^{\ast }\right) /\lambda
224: _{1}^{\ast }], \TCItag{4}
225: \end{eqnarray}%
226: respectively, where $\tau ^{\prime \prime }\equiv \sqrt{\kappa }\left( \tau
227: -\pi /4\right) ,\tau ^{\prime }\equiv \sqrt{\kappa }\left( \tau -\pi
228: /2\right) ,K_{0}\equiv \pi \sqrt{\kappa }\left\vert J_{1}\left( \lambda
229: _{0}^{\ast }\right) \right\vert /2,K_{1}\equiv \pi \sqrt{\kappa }\left\vert
230: J_{0}\left( \lambda _{1}^{\ast }\right) \right\vert ,T\equiv \pi \sqrt{%
231: \kappa },z\equiv x\pm \pi /2$ with the sign $+\left( -\right) $ when $%
232: J_{1}\left( \lambda _{0}^{\ast }\right) <\left( >\right) 0$, and where the
233: shift $x\rightarrow x\pm \pi $ is applied when $J_{0}\left( \lambda
234: _{1}^{\ast }\right) <0$. One sees that the description by means of the $%
235: \delta $-kicked rotors (3), (4) is almost perfect in the quantum regime
236: (i.e., for sufficiently large zeros), and that the shift $z\equiv x\pm \pi /2
237: $ in their spatial coordinates as well as the different values of the
238: amplitudes $K_{0},K_{1}$ help to understand why the approximate conservation
239: of the kinetic energy does not occur for the same set of initial conditions
240: at both kinds of zeros (cf. Figs. 3 and 4). Therefore, Eqs. (3) and (4) can
241: be reduced in the quantum regime to the standard maps [2]: $\overset{\_}{p}%
242: =p+K_{0}\sin z,\overset{\_}{z}=z+T\overset{\_}{p}$; $\overset{\_}{p}%
243: =p+K_{1}\sin x,\overset{\_}{x}=x+T\overset{\_}{p}$, per period $T$,
244: respectively. In such a case, the energy of a trajectory randomly fluctuates
245: inside the chaotic separatrix layer whose width is $\Delta E_{s,0}\left(
246: \lambda _{0}^{\ast },\kappa \right) \approx 64\pi \kappa ^{-1}\left\vert
247: J_{1}\left( \lambda _{0}^{\ast }\right) \right\vert ^{-1}\exp \left[ -\pi
248: \sqrt{2/\left( \kappa \left\vert J_{1}\left( \lambda _{0}^{\ast }\right)
249: \right\vert \right) }\right] $, $\Delta E_{s,1}\left( \lambda _{1}^{\ast
250: },\kappa \right) \approx 32\pi \kappa ^{-1}\left\vert J_{0}\left( \lambda
251: _{1}^{\ast }\right) \right\vert ^{-1}\exp \left( -\pi /\sqrt{\kappa
252: \left\vert J_{0}\left( \lambda _{1}^{\ast }\right) \right\vert }\right) $,
253: respectively. Note that these estimates are in excellent agreement with Eq.
254: (2). For example, for the experimental value $\kappa =0.36$ [9,11] one has $%
255: \Delta E_{s,0}\left( 8.6537,0.36\right) \simeq 0.00138,\Delta E_{s,1}\left(
256: 7.0156,0.36\right) \simeq 0.06574$ while $d\left( 8.6537,0.36\right) \simeq
257: 0,d\left( 7.0156,0.36\right) \simeq 0.06769$. In general, $\Delta
258: E_{s,0}\left( \lambda _{0}^{\ast },0.36\right) ,\Delta E_{s,1}\left( \lambda
259: _{1}^{\ast },0.36\right) $ are very small widths at sufficiently large
260: values of $\lambda _{0}^{\ast },\lambda _{1}^{\ast }$, which explains the
261: approximate conservation of the kinetic energy at such zeros for the
262: majority of the initial conditions.
263:
264: \textit{Conclusion.}$-$Now an interpretation of the above results in the
265: framework of Floquet theory is in order: They indicate that increasing the
266: strength of chaos implies decreasing $\tau _{E}$, i.e., the time interval in
267: which the Floquet states significantly contributing to the dynamics become
268: completely dephased. Thus, the maxima of the chaotic layer width could be
269: associated with corresponding minima of $\tau _{E}$. This means that when
270: the width in energy of the chaotic layer reaches a maximum as a function of
271: the modulation amplitude, the corresponding quantum momentum distribution
272: covers a minimal set of significantly populated Floquet states, and the
273: effect of DL is maximally enhanced. This picture could well explain some
274: previously discussed ways of breaking DL, such as deviation from strict
275: periodicity in the driving [18]. Indeed, it has been demonstrated that the
276: suppression of chaos [19] and the destruction of DL [18] by a secondary
277: irrational frequency are both very sensitive phenomena. It is also coherent
278: with the destruction of DL due to dissipation [20]: decreasing chaos by
279: increasing dissipation is a universal mechanism [2]. Thus, the present
280: theoretical approach to DL suggests that diverse methods for controlling
281: chaos [19] could also be useful in controlling DL. Finally, \textit{exact}
282: experimental realization of the standard $\delta $-kicked rotor by using
283: ultracold atoms interacting with a standing wave of laser light can be
284: achieved at the aforementioned suitable values of the modulation amplitude.
285:
286: The author acknowledges P. Bardroff \ and W. P. Schleich for kindly
287: providing the numerical data depicted in Fig. 1. This work was supported by
288: the Spanish MCyT and the European Regional Development Fund (FEDER) program
289: through project FIS2004-02475.
290:
291: \begin{thebibliography}{99}
292: \bibitem{1} L. E. Reichl, \textit{The Transition to Chaos in Conservative
293: Classical Systems: Quantum Manifestations} (Springer-Verlag, Berlin, 1992).
294:
295: \bibitem{2} A. L. Lichtenberg and M. A. Lieberman, \textit{Regular and
296: Chaotic Dynamics} (Springer-Verlag, Berlin, 1991).
297:
298: \bibitem{3} F. Haake, \textit{Quantum Signatures of Chaos} (Springer,
299: Berlin, 2001).
300:
301: \bibitem{4} G. P. Berman and G. M. Zaslavsky, Physica A \textbf{91}, 450
302: (1978).
303:
304: \bibitem{5} B. V. Chirikov, F. M. Izrailev, and D. L. Shepelyansky, Sov.
305: Sci. Rev. \textbf{2C}, 209 (1981).
306:
307: \bibitem{6} M. El Ghafar \textit{et al}., Phys. Rev. Lett. \textbf{78}, 4181
308: (1997).
309:
310: \bibitem{7} S. Fishman, D. R. Grempel, and R. E. Prange, Phys. Rev. Lett.
311: \textbf{49}, 509 (1982).
312:
313: \bibitem{8} R. Graham, M. Schlautmann, and P. Zoller, Phys. Rev. A \textbf{45%
314: }, R19 (1992).
315:
316: \bibitem{9} F. L. Moore, J. C. Robinson, C. Bharucha, P. E. Williams, and M.
317: G. Raizen, Phys. Rev. Lett. \textbf{73}, 2974 (1994).
318:
319: \bibitem{10} P. J. Bardroff, I. Bialinicki-Birula, D. S. Krahmer, G.
320: Kurizki, E. Mayr, P. Stifter, and W. P. Schleich, Phys. Rev. Lett. \textbf{74%
321: }, 3959 (1995).
322:
323: \bibitem{11} J. C. Robinson, C. Bharucha, F. L. Moore, R. Jahnke, G. A.
324: Georgakis, Q. Niu, M. G. Raizen, and B. Sundaram, Phys. Rev. Lett. \textbf{74%
325: }, 3963 (1995).
326:
327: \bibitem{12} For a recent review, see F. Saif, Phys. Rep. \textbf{419}, 207
328: (2005); \textit{Corrigendum}, Phys. Rep. \textbf{425}, 369 (2006).
329:
330: \bibitem{13} D. A. Wisniacki, E. Vergini, R. M. Benito, and F. Borondo,
331: Phys. Rev. Lett. \textbf{94}, 054101 (2005).
332:
333: \bibitem{14} R. Graham, M. Schlautmann, and D. L. Shepelyansky, Phys. Rev.
334: Lett. \textbf{67}, 255 (1991).
335:
336: \bibitem{15} V. I. Arnold, \textit{Geometrical Methods in the Theory of
337: Ordinary Differential Equations} (Springer-Verlag, New York, 1988).
338:
339: \bibitem{16} Dependence of DL on the initial momentum has been demonstrated
340: for a kicked rotor pulsed with unequal periods in T. Jonckheere, M. R.
341: Isherwood, and T. S. Monteiro, Phys. Rev. Lett. \textbf{91}, 253003 (2003).
342:
343: \bibitem{17} In particular, $\left( 2n/\lambda \right) J_{n}\left( \lambda
344: \right) =J_{n-1}\left( \lambda \right) +J_{n+1}\left( \lambda \right) ,2d%
345: \left[ J_{n}\left( \lambda \right) \right] /d\lambda =J_{n-1}\left( \lambda
346: \right) -J_{n+1}\left( \lambda \right) $, $n$ integer. See, e.g., M.
347: Abramowitz and I. Stegun, \textit{Handbook of Mathematical Functions}
348: (Dover, New York, 1970).
349:
350: \bibitem{18} G. Casati \textit{et al}., Phys. Lett. A \textbf{154}, 19
351: (1991); J. Ringot \textit{et al}., Phys. Rev. Lett. \textbf{85}, 2741
352: (2000); H. Lignier \textit{et al}., Phys. Rev. Lett. \textbf{95}, 234101
353: (2005).
354:
355: \bibitem{19} R. Chac\'{o}n, \textit{Control of Homoclinic Chaos by Weak
356: Periodic Perturbations} (World Scientific, London, 2005).
357:
358: \bibitem{20} T. Dittrich and R. Graham, Ann. Phys. \textbf{200}, 363 (1990);
359: B. G. Klappauf, W. H. Oskay, D. A. Steck, and M. G. Raizen, Phys. Rev. Lett.
360: \textbf{81}, 1203 (1998).
361: \end{thebibliography}
362:
363: \subsection{Figure Captions}
364:
365: \bigskip
366:
367: Figure 1. (a) Comparison between the classical [thick blue (black) line] and
368: quantum mechanical [red (gray) line] momentum distribution of atoms moving
369: in a phase-modulated standing light field for $\tau =320$ (cf. Ref.[10]).
370: Here, $\Delta p\equiv p_{RMS}/(2\hbar k)\equiv \left[ \left\langle
371: p^{2}\right\rangle -\left\langle p\right\rangle ^{2}\right] ^{1/2}/(2\hbar k)
372: $ is a normalized RMS momentum width. Also plotted for comparison is the
373: width of the chaotic layer [Eq. (2), thin black line] vs the amplitude $%
374: \lambda $. (b) Difference between the classical and quantum mechanical
375: momentum distributions $\Delta p_{C-Q}$ (green line), width of the chaotic
376: layer (black line), and Bessel functions $3\left\vert J_{1}\left( \lambda
377: \right) \right\vert $ (dashed line) and $3\left\vert J_{0}\left( \lambda
378: \right) \right\vert $ vs the amplitude $\lambda $.
379:
380: \bigskip
381:
382: Figure 2. Maximal LE corresponding to $\kappa =0.36$ and nine initial
383: conditions inside the initial well and their average (thick black line), $%
384: \Lambda ^{+}$, and the rescaled chaotic layer width [Eq. (2), red (gray)
385: line] as functions of the amplitude $\lambda $. The null value of $\Lambda
386: ^{+}$ for finite ranges of $\lambda $ around the zeros of $J_{1}\left(
387: \lambda \right) $ indicates quasi-periodic motion.
388:
389: \bigskip
390:
391: Figure 3. Grids of $75\times 75$ points in the $x(0)-\lambda $ plane for $%
392: \kappa =0.36$ and $p(0)=0$. Blank, yellow, red, and black regions indicate
393: that the maximal LE, $\Lambda ^{+}$, satisfies $\left\vert \Lambda
394: ^{+}\right\vert <10^{-3},\Lambda ^{+}\in \left[ 10^{-3},0.06\right] ,\Lambda
395: ^{+}\in (0.06,0.12],$ and $\Lambda ^{+}>0.12$, respectively. The zeros of $%
396: J_{0}\left( \lambda \right) $ are also indicated for comparison with the
397: regions where the kinetic energy is approximately conserved (blank regions).
398:
399: Figure 4. Grids of $75\times 75$ points in the $x(0)-\lambda $ plane for $%
400: \kappa =0.36$ and $\ p(0)=-1$. Blank, yellow, red, and black regions
401: indicate that the maximal LE, $\Lambda ^{+}$, satisfies $\left\vert \Lambda
402: ^{+}\right\vert <10^{-3},\Lambda ^{+}\in \left[ 10^{-3},0.06\right] ,\Lambda
403: ^{+}\in (0.06,0.12],$ and $\Lambda ^{+}>0.12$, respectively. The zeros of $%
404: J_{1}\left( \lambda \right) $ are also indicated for comparison with the
405: regions where the kinetic energy is approximately conserved (blank regions).
406:
407: \end{document}
408: